For years, Adobe has been pushing the envelope with its software. The Photoshop maker was a pioneer in moving its design software to a subscription model, a move that caused some short-term pain but ultimately boosted profit and sent the stock on a yearslong rally.
Even so, there were doubts earlier this year from some investors who feared that Adobe (ticker: ADBE) could be a net loser from artificial intelligence—that free applications from various vendors might hurt the company’s sales. In fact, the opposite seems true....
For years, Adobe has been pushing the envelope with its software. The Photoshop maker was a pioneer in moving its design software to a subscription model, a move that caused some short-term pain but ultimately boosted profit and sent the stock on a yearslong rally.
Even so, there were doubts earlier this year from some investors who feared that Adobe (ticker: ADBE) could be a net loser from artificial intelligence—that free applications from various vendors might hurt the company’s sales. In fact, the opposite seems true. Adobe is positioned to be one the biggest—and earliest—winners from the AI revolution.
Over the past few weeks, Adobe has made a series of announcements regarding its growing portfolio of generative-AI software. Aside from Microsoft (MSFT) and Alphabet (GOOGL), Adobe might be software’s most aggressive mover in revamping business for the age of AI.
That shouldn’t be surprising to anyone who has been watching Adobe. The company has become the leading provider of creative software, and an important player in marketing and document-management tools, like e-signature.
READ MORE
In a recent interview with Barron’s, Adobe CEO Shantanu Narayen made the case that generative AI is going to expand the number of people creating their own content. Rather than hurting Adobe’s business, he says that AI will boost interest and usage of Adobe’s creative software applications, like Photoshop, Illustrator, and Premiere Pro.
Until now, Adobe has been giving away its new tools on a new website called Firefly, where users can mess around for free with some astonishing (and easy to use) new photo-editing tools, along with a service that can create images from a simple text command, similar to the Dall-e app from ChatGPT creator OpenAI.
Adobe says that users have already generated more than 200,000 digital objects on Firefly, and 150 million more using a beta version of new tools for Photoshop. So far, Adobe has been giving away those tools to all comers. But that will soon change, positioning Adobe for a more profitable AI future.
Here are highlights from Barron’s recent conversation with Narayen, who has been CEO of Adobe since 2007:
On Adobe’s view of generative AI:
AI is not new to us. It already helps us with our mission to help people express their ideas, and to make our products accessible and affordable. The excitement right now is clearly about generative AI.
We’ve always thought that generative AI can serve as a launchpad to what people can do with our software—that’s where magic will happen. It will help people accomplish what they want to accomplish, and faster.
Adobe thinks about it in three layers. First, if we are going to use generative AI, we need to integrate it into existing products to make software more effective for customers.
Second, we’re creating fundamental foundational models, because we are the experts in this area, and we have the most passion for it. But our approach is not always the same. For imaging, with Firefly, we believed that we had the ability to create a fundamental foundational model, so we built one. For text, we thought ChatGPT already was the state of the art, and decided to leverage that model rather than creating our own.
In digital marketing, with PDFs, we previously had created a model that allows us to do a feature called “liquid mode,” where we can repurpose any PDF for a mobile device.
The third layer is data. We took a differentiated approach to training our models, using only images in Adobe Stock photo library, where we have the content under license, or in the public domain.
We concluded that Adobe had to assume responsibility not just for ourselves, but also for the industry, in terms of how people should be using data, and who owns the data.
On how AI compares to other tech trends:
This is going to be a profound technology inflection point, like mobile devices were, like the move toward the cloud was, like the advent of social media.
On the outlook for regulation of artificial intelligence:
This notion that “we’re going to have a moratorium on AI,” when the rest of the world may not, is not the right thing. We need to harness the power of AI, while understanding the ramifications. Prematurely regulating AI would be misguided.
On the early success of Firefly:
It has been one of our most successful betas—but there is more to come. We have an exciting road map of co-pilots as interfaces. Imagine the power of Firefly for all the different things we’re doing—fonts, vectors, colors, videos, when that’s integrated into each of our creative applications, like the generative fill feature we’ve introduced in beta in Photoshop.
On making money from Firefly and AI tools:
Firefly will be a separate service. It’s in a free beta right now. When it goes to general availability, we will still have a free service, but as recently announced, we will also have a subscription service with higher quality, higher usage caps, and different ways of using it.
Clients will be able to train their own models, with their own content, whether you are Coca-Cola or Barron’s. Some companies will want to tweak our model with their own content, but make that refined model accessible only inside the enterprise.
On finding ways to make AI commercially safe, while protecting intellectual property:
The moment we announced Firefly, we made it clear that this was designed to be commercially safe. With some of the other models, you can do things that clearly infringe copyrights or trademarks. With Firefly, you cannot do that.
And so the interest level, as you can imagine, has come from media companies, from [consumer-package goods] companies, from other companies. They all see the power of generative AI, but they want either the power of their data and their assets to be monetized exclusively by them, or they want to be compensated for it.
In some cases we have created custom models, where people have given us a subset of their data. And then, every single individual or knowledge worker at an enterprise can do content production, or campaign creation and marketing. Why wouldn’t that be just as valuable to every knowledge worker in an enterprise as an email application or a word-processing application?
On compensating artists that have contributed content for Adobe’s stock library:
The feedback to our plans to compensate developers for work created from their stock images has been incredibly positive. People feel like we’ve taken both innovation and the community into consideration.
Creators appreciate our desire to engage with them to come to a solution. Certainly, with new technologies there will always be some people who have a negative reaction. But I have been pleasantly surprised by how overwhelmingly positive the feedback to what we have done has been.
On how AI can expand Adobe’s total addressable market:
We’ve been unambiguous about that—it creates more TAM. When mobile devices came along, people were like, “Oh my God, how are people going to create content for that?” When the web came along, the question was, “Are people going to create more content?”
And now they are asking the same questions about AI. We are living in this world where every customer expects everything to be personalized—and that suggests that the amount of content that is going to be created is going to go up and to the right—no ifs, ands, or buts.
We talk about our TAM as $200 billion. We’re at $20 billion in annual revenue right now, so already we have a large opportunity. But as AI expands, with more people using this technology, that is an accelerant to our growth.
Write to Eric J. Savitz at eric.savitz@barrons.com